When detecting a face included in an input image by use of at least two inclinations of the face to be detected, a first detection process is performed if the inclination of the face to be detected corresponds to a first inclination where a relatively high face detection probability is expected, and a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process is performed if the inclination of the face to be detected is other than the first inclination. When an input image is scanned to enable the change in face detection processes according to the position of the face to be detected, a different detection process is used according to the face detection probability of the position of the face to be detected.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A face detection method for detecting a face included in an input image by use of at least two inclinations of the face to be detected, wherein a first detection process is performed, if the inclination of the face to be detected corresponds to a first inclination where a relatively high face detection probability is expected, otherwise a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process is performed, if the inclination of the face to be detected is other than the first inclination, wherein the method is performed using a processor or a computer.
2. The face detection method according to claim 1 , wherein the first inclination is an inclination substantially corresponding to the vertical or horizontal direction of the input image.
3. The face detection method according to claim 2 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
4. The face detection method according to claim 1 , wherein: two different types of detectors, a first detector for performing the first detection process, and a second detector for performing the second detection process, are provided in advance for each inclination of the face to be detected; and either the first or second detection process is performed by selecting and using either of the two different types of detectors according to the inclination of the face to be detected.
5. The face detection method according to claim 4 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
6. The face detection method according to claim 1 , wherein: each of the first and second detection processes performs detection while varying the position of the face to be detected on the input image every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the position of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the position of the face to be detected.
7. The face detection method according to claim 1 , wherein: each of the first and second detection processes performs detection while varying the size of the face to be detected every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the size of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the size of the face to be detected.
8. The face detection method according to claim 1 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
9. A face detection method in which an input image is scanned to enable the change in face detection processes according to the position of the face to be detected, wherein a first detection process is performed, if the position of the face to be detected corresponds to a first position where a relatively high face detection probability is expected, otherwise a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process is performed, if the position of the face to be detected is other than the first position, wherein the method is performed using a processor or a computer.
10. The face detection method according to claim 9 , wherein the first position is a position within a predetermined region of the input image including the center thereof.
11. The face detection method according to claim 10 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
12. The face detection method according to claim 9 , wherein: each of the first and second detection processes performs detection while varying the position of the face to be detected every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the position of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the position of the face to be detected.
13. The face detection method according to claim 9 , wherein: each of the first and second detection processes performs detection while varying the size of the face to be detected every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the size of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the size of the face to be detected.
14. The face detection method according to claim 9 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
15. A face detection apparatus for detecting a face included in an input image by use of at least two inclinations of the face to be detected, the apparatus comprising a detector module to perform a first detection process, if the inclination of the face to be detected corresponds to a first inclination where a relatively high face detection probability is expected, otherwise a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process, if the inclination of the face to be detected is other than the first inclination.
16. The face detection apparatus according to claim 15 , wherein the first inclination is an inclination substantially corresponding to the vertical or horizontal direction of the input image.
17. The face detection apparatus according to claim 16 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
18. The face detection apparatus according to claim 15 , wherein: the detection means includes two different types of detectors, a first detector for performing the first detection process, and a second detector for performing the second detection process; and either the first or second detection process is performed by selecting and using either of the two different types of detectors according to the inclination of the face to be detected.
19. The face detection apparatus according to claim 18 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
20. The face detection apparatus according to claim 15 , wherein: each of the first and second detection processes performs detection while varying the position of the face to be detected on the input image every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the position of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the position of the face to be detected.
21. The face detection apparatus according to claim 15 , wherein: each of the first and second detection processes performs detection while varying the size of the face to be detected every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the size of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the size of the face to be detected.
22. The face detection apparatus according to claim 15 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
23. A face detection apparatus that performs detection by scanning an input image and changing the position of the face to be detected on the input image, the apparatus comprising a detection module to perform a first detection process, if the position of the face to be detected corresponds to a first position where a relatively high face detection probability is expected, otherwise a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process, if the position of the face to be detected is other than the first position.
24. The face detection apparatus according to claim 23 , wherein the first position is a position within a predetermined region of the input image including the center thereof.
25. The face detection apparatus according to claim 24 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
26. The face detection apparatus according to claim 23 , wherein: each of the first and second detection processes performs detection while varying the position of the face to be detected every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the position of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the position of the face to be detected.
27. The face detection apparatus according to claim 23 , wherein: each of the first and second detection processes performs detection while varying the size of the face to be detected every time by a predetermined amount of variance; the first detection process is a detection process with a relatively small amount of variance in the size of the face to be detected; and the second detection process is a detection process with a relatively large amount of variance in the size of the face to be detected.
28. The face detection apparatus according to claim 23 , wherein: each of the first and second detection processes uses a discriminator that comprises a plurality of different types of weak classifiers connected in cascade for cutting out a partial image at the position of the face to be detected, calculating an index value that indicates the probability that the partial image is a face image based on the characteristic amounts of the image of the partial image, and discriminating whether the partial image is a face image based on a threshold for the index value, to discriminate whether the partial image is a face image; and the first detection process is a detection process with a relatively low threshold for the index value, and the second detection process is a detection process with a relatively high threshold for the index value.
29. A computer readable recording medium on which a program for causing a computer to function as a face detection apparatus stored thereon or embodied therein for detecting a face included in an input image by use of at least two inclinations of the face to be detected, wherein the program causes the computer to function as a detection means performing a first detection process if the inclination of the face to be detected corresponds to a first inclination where a relatively high face detection probability is expected, otherwise a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process, if the inclination of the face to be detected is other than the first inclination.
30. A computer readable recording medium on which a program for causing a computer to function as a face detection apparatus stored thereon or embodied therein in which an input image is scanned to enable the change in face detection processes according to the position of the face to be detected, wherein the program causes the computer to function as a detection means performing a first detection process if the position of the face to be detected corresponds to a first position where a relatively high face detection probability is expected, otherwise a second detection process having a relatively low face detection probability and a relatively short processing time compared with the first detection process, if the position of the face to be detected is other than the first position.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
October 31, 2006
October 5, 2010
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.